TLDR DevOps 2025-11-17
Python Performance 🐍, Distributed AI/ML at Scale ⚖️, AI Cyber Espionage 🥷
GitHub Copilot CLI 101: How to use GitHub Copilot from the command line (8 minute read)
GitHub Copilot CLI lets users interact with Copilot directly from the terminal to generate scripts, refactor code, run commands, and get explanations without leaving their workflow. It supports interactive and programmatic modes, integrates with MCP servers for custom capabilities, and helps automate tasks, maintain codebases, generate documentation, and prototype projects efficiently.
Finding the grain of sand in a heap of Salt (13 minute read)
Cloudflare reduced configuration management failures by over 5% through a new infrastructure that helps find the root cause of Salt failures on servers. The new system correlates failures with git commits, external service failures, and ad hoc releases, which has reduced software release delays and toil for SRE. A custom Salt execution module called "Salt Blame" pulls local job history, CM inputs, and GitHub repo commit history together.
Valdi (GitHub Repo)
Snap's Valdi, a cross-platform UI framework used in its production apps for 8 years, has been released in beta. The framework compiles declarative TypeScript components into platform-native views for iOS, Android, and macOS, avoiding web views and JavaScript bridges to improve performance.
TOON (GitHub Repo)
Token-Oriented Object Notation (TOON) is a compact and human-readable JSON alternative that minimizes tokens for LLM prompts by combining YAML's indentation with CSV-style tabular arrays. Benchmarks reveal that TOON achieves 73.9% accuracy, outperforming JSON's 69.7%, while using 39.6% fewer tokens. A TypeScript SDK is available, and community tools have been created for token comparison, format conversion, and validation.
Powering Distributed AI/ML at Scale with Azure and Anyscale (5 minute read)
Microsoft and Anyscale announced a partnership to bring Anyscale's managed Ray service to Azure as a native offering in private preview, combining Ray's open-source distributed computing with Azure's enterprise infrastructure. The integration enables developers to scale Python-based AI and ML workloads efficiently using Anyscale Runtime on Azure Kubernetes Service with unified governance, security, and scalability.
Eliminate unnecessary costs in your Amazon S3 buckets with Datadog Storage Management (5 minute read)
Datadog Storage Management helps organizations reduce Amazon S3 costs by providing prefix-level visibility into which teams, services, and datasets drive storage use. It identifies cold data, recommends cost-saving actions such as transitioning to cheaper tiers or consolidating small files, and integrates with workflow tools to streamline optimization decisions.
Disrupting the first reported AI-orchestrated cyber espionage campaign (5 minute read)
A Chinese state-sponsored group manipulated the Claude Code AI tool in mid-September 2025 to infiltrate roughly 30 global targets, marking what is believed to be the first documented case of a large-scale cyberattack executed without substantial human intervention. The AI was used to perform 80-90% of the campaign, analyzing target systems, producing exploit code, and scanning stolen data at speeds impossible for human hackers. It is advised that security teams should experiment with applying AI for defense, as this attack demonstrates how threat actors are adapting their operations to exploit today's most advanced AI capabilities.
Get our free daily newsletter with curated tools 💻, trends 📈, and insights 💡, for DevOps Engineers 👨💻
Join 340,000 readers for
one daily email